Please find below my April 4th effort (so before cessation of hostilities announced on April 5th) to update and expand my initial assessment of the conflict between Armenia and Azerbaijan.
While initial clashes, which began in the night from April 1 to April 2 along the western part of the line of contact (LOC), which separates Armenian and Azeri forces in Karabakh, the subsequent hostilities quickly spread as far as the most northern part of the border between Armenia proper and Azerbaijan in what became the bloodiest and heaviest episode of hostilities since the 1994 ceasefire. Such proliferation of hostilities, and, more important, use of multiple-launch rockets systems, heavy flame-throwing systems, artillery and attack helicopters made me assess the probability of a full-blown war as of April 4th at 30% in the subsequent two weeks.
That said, I continue to believe that it is unlikely that Azeris, which appear to have initiated the fighting, planned such a major escalation. Both Aliev and Sargsyan were in the air, flying back from NSS in DC, when the clashes erupted. You normally don’t let your army start a full-blown inter-state war, when the head of the state is traveling abroad. Rather Azeris may have planned a larger-scale, but still routine provocation to remind the international community that the conflict is not frozen beyond reheating, try coercing Armenia and Karabakh into concessions in the peace talks, which have stalled for years, and, perhaps seize a “strategic” height or two to claim a small victory. Instead they saw the situation escalate beyond original plans. If Azeris did indeed plan to attack Armenians along wide parts of LOC, they would have done it at once (initiating a main strike in place to distract Armenians and delivering it somewhere else within hours). Also, after having seized several heights in Karabakh and raided some settlements, Azeri MoD would not have announced that it began to observe a ceasefire on April 3rd and would have attempted on that day to advance further. That announcement was probably an attempt to secure the gains, but was ignored by Armenians who understand that they need to regain what’s been lost or Azeris would go even deeper next time.
Of course, while it is less probable, but it is not impossible that Azeris had planned to attack Armenians not just in Karabakh, but also along the proper border of Armenia and Azerbaijan all along. If true, an April 4th report Azeri defense minister Gasanov is personally commanding 1st Army Corps’ operations could be prove that the wide-scale hostilities had been planned along. However, it might be also be a sign that Azeri leadership has not been planning a full-scale war along LOC. A defense minister doesn’t assume command of one corps if the battle plan calls for coordinated operations of several corps or entire armed forces. Some posited that one possible reason why Aliev may want a full-blown war is that he needs a “great” victory to divert attention from Azerbaijan’s oil-price-induced economic woes, which already caused some social protests in the region. I don’t think, however, that it is the case. Azeris still have plenty of wealth to ride out the storm, unless, of course, it continues to rage on.
Regardless of initial intentions, both sides have climbed a step or two on the escalation ladder and may find it easier to keep climbing up, then descending. Even if the current hostilities do not escalate into a full-blown war, the latter’ probability has significantly increased.
For one, the April 2nd has created a precedent for use of yet heavier and more sophisticated weaponry. Now that Grad Multiple Launch Rocket Systems and Heavy Flamethrower System TOS-1s have been used along with Mi-35s, repetition of use of these systems is permissible. Therefore, when the next provocation occurs, commanders on both sides will have less doubts about employing large-caliber multiple rocket launch systems and gunships, especially if in “fog of hostilities” they come to conclude that the other side’s use of such weapons is the first stage of a full-blown major offensive. A provocation that begins with mortar fire has low chances of quickly escalating into a war even if leaders of neither side desire such a war. A provocation that begins with use of large-caliber multiple rocket launch systems and gunships has significantly higher chances of leading to an accidental war because of casualties it can cause. If Azeris did initiate the April 2nd clashes (and evidence is strong that they did), then this may also increase probability of a full-blown war because it may lead to hardening Armenians’ negotiating position.
Azeris’ initiation of what has turned out to be deadliest clash since 1994 also gives the Armenian side serious reasons to even more distrust proposals for a step-by-step resolution, in which Armenians first give up some of the land they control in exchange for Azeris’ commitment to honor results of a new self-determination referendum in Karabakh. Eliminating of the step-by-step approach, which I personally have had qualms about, significantly reduces what negotiation professionals describe as ZOPA (zone of possible agreement) between Armenians and Azeris, increasing the attractiveness of war as BATNA (best alternative to negotiated agreement) to the Azeri side. Also, memories of Azeri soldiers’ atrocities, which include alleged beheading of an Armenian soldier and cutting off ears of executed civilians, will not fade away easily, making it more difficult for Armenians to accept any compromises, which are needed on both sides, to make the peace deal a reality.
The latest clashes once more underscore how dangerously wrong are hopes, which some leaders nurture, that they can up the ante a bit for purposes of signaling resolve to the opposite side or coercing the latter into concession, and then control/manage the resultant escalation. Serzh Sargsyan appears to understand this risk – hence he warned on April 4th thay “further escalation of military actions may result in unpredictable&irreversible consequences,including all-out war.” But perhaps, not all decision-makers on both sides realize that SOPs, which the conflicting sides’ militaries begin to implement, following the ‘up-the-ante’ orders, may lead to escalation of a stand-off into a full-blown war against these leaders’ wishes, and so can accidents. Both U.S. and Soviet leaders learned those and other lessons after living through the Cuban Missile Crisis and we outlined seven of these lessons for the Karabakh conflict in a 2012 Belfer Center paper, including the illusory natures of hopes that escalation can be always firmly controlled. The paper ended with the following conclusion: “Institutionalizing lessons of the Cuban missile crisis would help leaders on both sides of the Karabakh conflict to avert an ‘accidental’ devastating war. If, of course, they wish to avoid it.” Maxim Yusin of Kommersant believes that leaders on both sides presently wish to avoid resumption of the war. Indeed, Armenian leaders obviously are content with what Armenians have gained in the war that ended in 1994 while Ilham Aliev probably realizes that there’s no guarantee that Azeris will win the new war, but that there is a strong probability that a defeat may lead to implosion of his regime. However, Armenian and Azeri leaders’ perceived aversion to resumption of full-blown war may change due to a number of factors. For instance, protracted decline of oil prices leaves Aliev facing protests so serious that he would come to view probability of losing his seat because of such protests greater than probability of losing a war over Karabakh, and then chances are that he may attempt such a war. A significant decline of one of the sides’ power relative to the other may also affect the cost-benefit analysis that leaders on both sides engage in as they weigh whether to go to war. 
A full-blown war between Armenia and Azerbaijan will have devastating consequences not only for these two post-Soviet republics, but also for stability of the region, in which U.S., Russia, EU, Turkey and Iran have stakes. For one, it can bring NATO member Turkey, which unconditionally backs their Azeri brothers, and Russia, which is Armenia’s ally in the Collective Security Treaty Organization (CSTO) into a war, which can begin as a proxy war, but can end up being a direct war between Russia and Turkey. Turkey has already been criticized for failing to protect interests of Crimean Tatars and stop Russia from bombing Turkomans in Syria and it feels much stronger about Azeris, with whom they share ethnicity and language. Such a war will also force Russia to take sides between its CSTO ally Armenia and what RF MFA has described as “friendly” Azerbaijan, therefore, undermining Moscow’s efforts to keep both within what RF MFA describes as a zone of Russia’s “privileged interests”. Russia has a lot at stake, including its reputation of a guarantor of security of its allies and reputation of CSTO as a viable military bloc, along with its desire to remain the strongest player in the South Caucasus. Some have speculated that the hostilities were encouraged by Russia, which wants to deploy peacekeeping troops along LOC to boost leverage over Azerbaijan or even let Putin score another foreign policy victory to prevent decline of popularity at home. However, even though his KGB evaluators described Putin as having “lower sense of danger,” and, therefore, more prone to take risks, I don’t think it is a very plausible theory, given that he already has his hands full with two conflicts (Syria and Turkey), which need to be resolved on terms that accommodate Russia’s wishes. Nor would Iran want resumption of war in Karabakh, given Teheran’s delicate effort to maintain good relations with Armenia without alienating its sizeable Azeri minority. If Azeri forces begin to retreat in such a war, it will send hundreds of thousands of Azeris seek refuge across the Azeri-Iranian border as it was the case during the (first and hopefully the last) Karabakh war. If Armenians began to lose, then Iran would see its rival – Turkey – expand its influence in the South Caucasus. U.S. clearly too has important (but not vital) interests at stake in the region, given Azerbaijan’s supplies to the world oil market, which Armenia can disrupt with ballistic missile strikes. U.S. also is member of OSCE’s three-strong Minsk Group, which is responsible for mediation of the Karabakh conflict (others are France and Russia). There is n estimated total of 800,000 — 1,500,000 Americans citizens of Armenian ethnicity .
I think it is time the presidents of U.S. and Russia, which both leverage vis-a-vis Armenia and Azerbaijan, step up and makes a coordinated push with fellow leaders of the Minsk Group before it escalates into a full-blown war, in which none of major stake-holders realistically stand anything substantial to gain.
 Such analysis applies, however, only if we view both Armenia, Karabakh and Azerbaijan as unitary rational actors acting in their national interest in what would conform to behavior of Model I actors in Graham Allison’s study of decision-making. However, there are also Model II and Model III and that’s where organizational interests and internal government politics come into play in ways that may enhance probability of full-blown war against heads of states’ wishes.
 Russian officials may also now be pondering whether their country can keep supplying arms to both Armenia (got 0.2bn of Russian arms) and “friendly” Azerbaijan under the pretext that is done “taking into account the need to maintain balance of forces in the region.”